CRI researchers Marc Santolini and Bastian Greshake Tzovaras explore how the tweets on self-reported COVID19 symptoms can help predict future pandemic waves, and, more generally, the rise and fall of the disease. With the help of Naila El Haouari (intern at CRI) and Samuel Fraiberger (NYU & MIT MediaLab), they scanned public tweets from the Paris region and filtered them by symptom keywords. Although this already seems to give a good correlation between the number of tweets and disease progression, the automatic filtering is very crude. In general, people don't just tweet about symptoms when they're getting sick, but also about past times when they've been ill, or when they're talking about general news.
To filter out these false positives, Marc & Bastian need your help to see which tweets describe an acute symptom and which do not ? Your contribution will have a direct impact! Thanks to this additional filtering, they will not only be able to refine their current analyses, but also train the artificial intelligence to create better automated filtering strategies.
Your help doesn't have to be limited to annotating tweets. As CRI is dedicated to open science, all data and analysis methods are shared publicly. So you can download tweets and annotations yourself to reproduce results and improve analyses.
Visit the project website to find out more and start contributing: http://covid-twitter.thecommons.science/.
